caretFitQtl: Fit with qtl for caret
In timflutre/rutilstimflutre: Timothee Flutre's personal R code

caretFitQtl

R Documentation

Fit with qtl for caret

Description

Fits the simple or multiple interval mapping model for usage with caret. First, the genotypes in the JoinMap format are saved into a temporary file. Second, a cross object is made. Third, SIMQTL or MIMQTL is run (calc.genoprob, scanone or scantwo). Fourth, allelic effects are estimated.

Usage

caretFitQtl(
  x,
  y,
  wts,
  param,
  lev,
  last,
  weights,
  classProbs,
  genmap,
  cross.geno.prob = NULL,
  phase,
  tuneThreshold = TRUE,
  alpha = c(0.05, NA),
  nperm = 10,
  pop.type = "CP",
  QTLmethod = "SIM",
  p2d = "",
  plot = FALSE,
  QTL_position = NULL,
  nb.cores = parallel::detectCores() - 2,
  verbose = 0
)

Arguments

`x`	current predictors used to fit the model, i.e genotypic matrix in Joinmap format with genotypes in rows and markers in columns
`y`	current outcome used to fit the model, i.e trait to study in numeric format with genotype names
`wts`	optional instance weights
`param`	data frame with current tuning parameter values
`lev`	class levels of the outcome (or NULL in regression)
`last`	logical for whether the current fit is the final fit
`weights`	?
`classProbs`	logical for whether class probabilities should be computed
`genmap`	genetic map in data frame with columns linkage.group, genetic.distance and locus
`cross.geno.prob`	list containing genotypic class probabilities calculated from calc.genoprob. If not passed (NULL), prediction is done with allelic effects at the closest marker.
`phase`	vector of character (length nb of markers) in the format '–'
`tuneThreshold`	logical, if TRUE (default), the threshold is chosen for minimizing prediction error with a criterion, else threshold is chosen by permutations for minimizing False Positive error in QTL detection.
`alpha`	numeric vector with first element corresponding to alpha rate error for chosing LOD significance threshold after permutations, and the 2nd element is a threshold applied after linear regression with selected markers, if NA, no selection is applied.
`nperm`	number of permutations to be done is tuneThreshold=FALSE
`pop.type`	population type in the JoinMap format, only "CP" is handled here
`QTLmethod`	character (SIM/MIM)
`p2d`	path to directory with results of `scantwo` (passed to MIMQTL)
`plot`	logical, default is FALSE. If method is MIM, 2 arguments are needed.
`QTL_position`	data frame with columns linkage.group, genetic.distance and locus to plot abline at simulated QTL position
`nb.cores`	number of cores (passed to MIMQTL)
`verbose`	verbosity level (0/1/2)